Optimization Research on DNA Sequence Alignment Algorithm Based on Deep Learning: Innovative Approach and Application

The DNA sequence alignment algorithms are facing challenges due to the booming genome sequencing technologies which provide large volumes of data. This research proposes HybridAlign, a novel sequence alignment algorithm. HybridAlign utilizes MinHash, k-mer indexing, and suffix arrays, and applies attention techniques and dynamic programming. The results show that HybridAlign significantly improves the efficiency of sequence alignment. Tests with 1000 genomic project datasets show that HybridAlign has considerable improvements in both precision and speed over the traditional methods. The rate of detection is 15% higher, while the computation time is 40% lower while processing complex variant sequences. Thus, the research provides a useful and reliable method for exploring large genomic data.